Releases: Coherent-Partners/spark-python-sdk
Releases · Coherent-Partners/spark-python-sdk
Release list
0.3.2
0.3.1
- Add support for decoding and validating JSON web tokens (JWT)
- Introduce
Spark.JwtConfigclass to simplify the creation of aSpark.Clientinstance from a JWT - Add ability to check the existence of a folder or service in Spark
- Update documentation and examples to reflect changes
- Add tests for
Spark.JwtConfigclass
0.3.0
- Describe import and export jobs across a tenant
- Introduce additional metadata when creating a service
- Fix bug in
Batches.create(...):max_input_sizeparameter was renamed tomax_input_in_mbmax_output_sizeparameter was renamed tomax_output_in_mb- add fail-safe
extrasto support additional parameters in the request body
- Remove
dataclassesdependency (supported as built-in from Python 3.7+) - BREAKING CHANGES:
- Refactor synchronous clients to use context manager pattern
- Add asynchronous client and resources for both SaaS and Hybrid Runner APIs
- Add basic tests for asynchronous client and resources
- Add migration guide for 0.2.x to 0.3.x (see migration guide)
- Update documentation and tests to reflect changes
0.2.3
0.2.2
0.2.1
- Add status (v1.46.0+ only), validation and metadata APIs for WASM services
- Add examples for batch status check:
examples/usecases/async_batch/check_status.py - Update documentation and references for SDK and WASM APIs
- Fix bug in
Spark.history.rehydrate()method (usePOSTinstead ofGET)
0.2.0
- Add support for Transform documents API (including Transform validator)
- Add health check functionality (for Spark environment only)
- Update rehydration of service execution to support new (and legacy) format
- WASM download is now available using different URI locators
- Update SDK documentation to reflect changes
0.2.0-rc.1
- Add support for Transform documents API (including Transform validator)
- Add health check functionality (for Spark environment only)
- Update rehydration of service execution to support new (and legacy) format
- WASM download is now available using different Uri locators
- Update SDK documentation to reflect changes
0.1.12
- Add support for Folder API
- Update
create_chunksmethod to handle columnar format only (headers + inputs) - Remove
Nonevalues from the request payload (e.g.,Services.execute()) - Support extra metadata in the request payload (e.g.,
Services.execute(extras={...})) - Update documentation and examples